ace-build

Pass

Audited by Gen Agent Trust Hub on Mar 2, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes Python scripts (scaffold.py, build.py) that perform filesystem operations and directory management. These are intended to be executed by the user or agent to manage the skill lifecycle.
  • [COMMAND_EXECUTION]: Scripts use dynamic path manipulation (sys.path.insert) to locate and import functions from the local engine directory (src.engine). This is a common pattern for local tool development to avoid complex package installation requirements.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest user-provided markdown and text to generate new skill files. This creates a surface where untrusted data is integrated into the workspace.
  • Ingestion points: User-provided descriptions, markdown, and prompts as defined in content/shaping-process.md and AGENTS.md.
  • Boundary markers: None explicitly implemented in the script wrappers to separate user input from the build logic.
  • Capability inventory: Filesystem write access and directory creation via delegated functions in the project's engine.
  • Sanitization: No explicit sanitization or validation of input content is performed within the provided script entry points.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 2, 2026, 12:38 AM